Dolastatin 15, an active peptide isolated from the sea hare Dolabella auricularia (G. R. Pettit et al., J. Org. Chem. 54 (1989), 6005) and structurally related synthetic peptides which are described in WO 93-23 424 are very promising novel agents, some of which are undergoing clinical testing. Since isolation from the natural source (6.2 mg from 1600 kg of sea hare) is out of the question, interest is directed at suitable synthetic processes which make the agents available in sufficient quantity and purity on the industrial scale.
Two processes have been described for preparing dolastatin 15: the elegant synthesis of G. R. Pettit et al. (J. Am. Chem. Soc. 113 (1991), 6692 and Tetrahedron 50 (1994), 12097) prepares dolastatin 15 starting from proline methyl ester hydrochloride (scheme 1). ##STR3##
However, the process detailed above has the following disadvantages:
1. The starting compound, proline methyl ester hydrochloride, is extremely hygroscopic. It must be prepared with careful exclusion of moisture otherwise the crystalline material deliquesces with partial ester hydrolysis. This makes industrial preparation difficult.
2. The dipeptide VIII is prone to cyclize to the diketopiperazine XIII: ##STR4## This cyclization leads to only small losses of yield in the laboratory but interferes considerably with the preparation of larger amounts of substance on the industrial scale.
3. Methyl esters are used to prepare each of VIII, IX and X. With the aqueous workups which are necessary at these stages, partial hydrolysis of these esters to carboxylic acids occurs. This side reaction also increases in importance on scale-up because the times during which the product is in contact with water increase when the reaction is scaled up.
4. The depsipeptide unit XII is also sensitive to hydrolysis, and it comes into contact with water twice during the isolation and preparation of the final product. The longer contact times in the industrial process lead to losses of product in this respect too.
A second process for preparing dolastatin 15 is described by Poncet et al. (Tetrahedron 48, 20, 4115-4112) starting from the tert-butyl ester of proline (scheme 2). ##STR5## The hydrochloride of proline tert-butyl ester is less hygroscopic than the methyl ester used in scheme 1. The dipeptide XV is somewhat less prone to form diketopiperazine than is the methyl ester VIII. The costly unit XX, which is elaborate to prepare, is employed at a later stage in the synthesis so that less of this compound is used.
The disadvantage of the Poncet process (scheme 2) is that the tert-butyl ester is more complicated to prepare than is the methyl ester. The cleavage of this ester with trifluoroacetic acid is likewise complicated, resulting in a flammable, explosive gas and fluorine-containing waste which is difficult to dispose of.
In addition, after the linkage of XIX and XX to give dolastatin 15, the Z radical must be replaced by two methyl groups (scheme 3). In this operation there is loss of a further 20% of the valuable material at a relatively late stage in the synthesis. ##STR6## The Pettit and Poncet processes can also be used to prepare numerous agents of WO 93/23.424 which are structurally related to dolastatin 15. For example, the tetrapeptide acid XI can be linked to proline benzylamide to give agent No. 234 from WO 93/23.424 (scheme 4). ##STR7## The pentapeptide acid XIX from scheme 2 can be reacted in a similar manner with a dipeptide to produce the heptapeptide XXII, from which agent No. 1 from WO 93/23.424 can then be prepared (scheme 5). ##STR8## The problems described above adversely affect the yields and the industrial implementability of the peptide synthesis for preparing agents via XIX and XI.